MLOps – Applying DevOps Practices to Machine Learning Workloads
Who is this presentation for?Data Scientists, DevOps, Automation Engineers
Description of the presentation:
In Gartner’s 2017 report “Predicts 2017: Analytics Strategy and Technology”, they estimate that by 2020 40% of all data science tasks will be automated. While the movement towards bringing the full potential of AI to the market faster is promising, there is a balance between automation and quality that needs to be maintained. Applying DevOps practices to machine learning workloads not only brings models to the market faster but also maintains the quality and integrity of those models. This presentation will focus on how to apply DevOps practices to machine learning workloads as well as demonstrate a CI/CD pipeline using a managed machine learning service and cloud based development services.
While the concept of applying DevOps practices to Machine Learning workloads is not necessarily novel, it is still very immature in practice. Business stakeholders are frequently looking for guidance on how to increase time to market and ensure models make it out of the lab. It’s estimated that a large number of models created never get deployed to production. To be able to realize the full potential of AI, it’s critical to ensure the models that get developed have a full path to production. That full path must include quality gates and varying levels of model evaluation using practices that are unique to machine learning workloads.
This technical talk will focus on the following areas to ensure audience members are able to take away key practices that can be applied to their existing workloads:
- DevOps 101: To understand MLOPs it’s important to have a base foundation of DevOps practices including what DevOps is and what DevOps is not.
- MLOps: Because the area of applying DevOps practices to machine learning workloads is still very immature, we will define the practices behind MLOps as well as step-by-step guidance on how to apply DevOps practices to machine learning workloads. The end result will be a Continuous Integration/Continuous Delivery (CI/CD) pipeline targeted for machine learning.
- Demonstrable CI/CD pipeline utilizing AWS services
The demonstrable CI/CD pipeline will utilize AWS services; however, the presentation will be largely agnostic of specific technologies so that the practices discussed can be applied regardless of technology or platform.
Prerequisite knowledgeBase knowledge of how a model is created is beneficial but not required.
What you'll learn
Amazon Web Services
Sireesha is a Specialist Solutions Architect,MachineLearning/AI Services at Amazon Web Services (AWS). She provides guidance to AWS customers on their ML/AI workloads. While working full-time, Sireesha earned her Ph.D in May 2013 and Post Doctorate in 2015 from University of Colorado, Colorado Springs. Her Ph.D thesis is, “Multi-tier Internet Service Management using Statistical Learning Techniques (https://dspace.library.colostate.edu/bitstream/handle/10976/264/CUCS2013100001ETDSPECS.pdf?sequence=1.)”. She led the Colorado University team in winning and successfully completing a 2-year research grant from Air Force Research Lab on “Autonomous Job Scheduling in Unmanned Aerial Vehicles”. She is an experienced public speaker and has presented research papers at International Conferences: CoSAC: Coordinated Session-Based Admission Control for Multi-Tier Internet Applications (https://www.researchgate.net/publication/221092402_CoSAC_Coordinated_Session-Based_Admission_Control_for_Multi-Tier_Internet_Applications) at IEEE Int’l Conf. on Computer Communications and Networks (ICCCN), 2009; Regression Based Multi-tier Resource Provisioning for Session Slowdown Guarantees (https://www.researchgate.net/publication/220780958_Regression_Based_Multi-tier_Resource_Provisioning_for_Session_Slowdown_Guarantees) at IEEE Int’l Conf. on Performance, Computing and Communications (IPCCC), 2010. She also published technical articles : Coordinated session-based admission control with statistical learning for multi-tier internet applications (https://www.researchgate.net/publication/222549520_Coordinated_session-based_admission_control_with_statistical_learning_for_multi-tier_internet_applications) in Journal of Network and Computer Applications (JNCA);Regression-based resource provisioning for session slowdown guarantee in multi-tier Internet servers (https://www.researchgate.net/publication/220379377_Regression-based_resource_provisioning_for_session_slowdown_guarantee_in_multi-tier_Internet_servers) and Multi-tier Service Differentiation: Coordinated Resource Provisioning and Admission Control (https://www.researchgate.net/publication/260042453_Multi-tier_Service_Differentiation_Coordinated_Resource_Provisioning_and_Admission_Control) in Journal of Parallel and Distributed Computing (JPDC)
Amazon Web Services
Shelbee Eigenbrode is a Solutions Architect at Amazon Web Services (AWS). Her current areas of depth include DevOps combined with Machine Learning/Artificial Intelligence. She has been in technology for 22 years spanning multiple roles and technologies. She spent 20+ years at IBM and joined AWS in November of 2018 She is a published author, blogger/vlogger evangelizing DevOps practices with a passion for driving rapid innovation and optimization at scale. In 2016, she won the DevOps Dozen Blog of the year demonstrating what DevOps Is Not. With over 26 patents granted across various technology domains, her passion for continuous innovation combined with a love of all things data has recently turned her focus to the field of Data Science. Combining her backgrounds in Data, DevOps and Machine Learning, her current passion is to help customers not only embrace data science but also to ensure all data models have a path to being utilized. She also aims to put ML is the hands of developers and customers that are not classically trained data scientists. Over her career, she has held senior leadership positions and has a passion for mentorship.
Amazon Web Services
Randy is a solutions architect at AWS, with over 20 years of experience in enterprise software architecture. He’s worked heavily in DevOps in the past, and currently focuses on Analytics and Machine Learning.
Leave a Comment or Question
Help us make this conference the best it can be for you. Have questions you'd like this speaker to address? Suggestions for issues that deserve extra attention? Feedback that you'd like to share with the speaker and other attendees?
Join the conversation here (requires login)
For conference registration information and customer service
For more information on community discounts and trade opportunities with O’Reilly conferences
For information on exhibiting or sponsoring a conference
View a complete list of Strata Data Conference contacts